In recent years, proper evaluation and management of habitat connectivity have been needed in fragmented landscapes. Floodplain is one of the most threatened ecosystems, and the floodplain pond is an ecosystem which has been degraded by intensive human land-use. The graph theory has been reported as a simple and powerful tool for analyzing habitat connectivity. We tested the effectiveness of this approach in aquatic ecosystems. In our study, as explanatory variables, we used pond-area and the habitat availability index varying with the pond-area and its position in the pond network. We clarified the strong variables that explain fish species richness.

We studied 24 ponds distributed over an altered floodplain, northern Japan. Fish sampling was conducted in 2011. We calculated habitat availability index of study ponds at the landscape scale by using existing landscape connectivity metrics. As a result, the index had a higher correlation with fish species richness than pond area did. This result indicated the significance of habitat configurations such as stepping stone effects for fish diversity. Among 9 native species captured in our study sites, habitat connectivity was particularly important for 2 species. Our study showed the applicability of the graph theory to freshwater ecosystem, and the results may greatly contribute the conservation of pond networks.